My in-room PB13 measurements + interpretation

**jvanhambelgium** · **07-31-08, 01:27 AM**

Hi,

Yesterday I finally got some time to measure around my PB13.
The most important reason was I wanted to check if the SVS was positioned well against the wall. (since I'm about to construct a piece of furniture that would make the current position of the sub rather permanent...)

Basically, sound quality is good (I think) through my Onkyo 875 with a Audessey engaged. I just wanted to "see" what Audessey would / would not have to compensate for...

After doing all the calibrations (soundcard + dB-meter) , loading the RadioShack file etc I placed the RS-meter on the "sweet spot" on the couch.(in the area where also the Audessey measurements are taken)
RS-meter was positioned at about "ear level" and flat (not at 45° or something)

Look below for some graphs.
Absolute dB-values may be not super accurate, but I'm more interested in the "shape" because I want to see what my large 8000cubic feet space does to the spectrum using a single PB13.

First measurement taken direct on the PB13 - bypassed Onkyo+Audessey (2 same graphs, different scale)

Second measurement taken through the whole chain (including Audessey)

Now, to ME it doesn't look that bad.
My PB13 is tuned to 15Hz so you see the steep rise @ 15Hz
Around 100Hz it goes down rapidly (Onkyo X-over value, 90Hz)

These graphs do not represent maximum output in any way, but I think the "shape" should be valid even if I'm some dB's off due to calibration-mismatch or whatever.

What do the more experienced user say about such response-graphs ??

Thanks & regards,
Jan

**Sonnie** · **07-31-08, 07:24 AM**

Hopefully you saved all of those measurements. The graphs size and axis make it difficult to compare to the norm.

Quote:

Ensure the graph is 800 pixels (or less) in width and that you have used a vertical scale of 45dB-105dB and a horizontal scale of 15Hz-200Hz.

**brucek** · **07-31-08, 07:31 AM**

I'd like to see two graphs using the standard axis we always use here. (I think you've nicely demonstrated how an exaggerated large or small axis can completely change how a response appears).

Can you change your axis to vertical = 45dB-105dB and horizontal = 15Hz-200Hz.

The two plots I would like to see are the final plot using those axis of frequency response and waterfall (log).

Use a target of 75dB.

brucek

**jvanhambelgium** · **07-31-08, 07:39 AM**

I've added a picture with the 2 curves combined and settings for the dB-scales

The global picture should be :

Green Curve : connected directly on PB13 (bypass Onkyo/Audessey) without any filters/X-overs on the PB13 used
Brown Curve : connected on AUX2 input on the Onkyo and processed by Audessey filters)
(I've "recalibated" when switching from direct to Onkyo-linkup but there might be a small difference between the 2 ...should not affect curves I guess...)

Does this look reasonably good from a response-perspective ? Me as an amateur would think it is not too bad without any major bass peaks / suckouts

Audessey seems to be something at least ;-)

**brucek** · **07-31-08, 07:51 AM**

Yeah, other than the tuning peak at ~16Hz, it's quite well behaved in response.

Can you do a measure (brown curve conditions) at 75dB and also include a log waterfall (using the same axis).

brucek

**JohnM** · **07-31-08, 11:44 AM**

Are you sure there's no smoothing applied to the brown curve? Looks like it has been smoothed to me.

**jvanhambelgium** · **07-31-08, 01:53 PM**

Whoops ... indeed the smoothing was activated on that one curve...

Below the composed graph without smoothing applied...

**brucek** · **07-31-08, 02:52 PM**

Ahhh OK. Not so great. You might want to see if other locations are a bit smoother.

You've also avoided showing me a waterfall. Is there a reason?

brucek

**jvanhambelgium** · **07-31-08, 06:48 PM**

Hi bruce,

For the waterfall, I just don't know what the best settings are ... the graph doesn't look that "usefull" to me...

Without audessey :

Through the whole chain:

**brucek** · **07-31-08, 07:09 PM**

Quote:

For the waterfall, I just don't know what the best settings are ... the graph doesn't look that "usefull" to me...

They're quite useful. They show you what happens over time, after the measurement sweep.

You really need to lower your target to 75dB, then the normal scaling of 45dB-105dB places the target in the middle of the plot.

If you move the waterfall slider from 30 back to 1, you will see your frequency response curve. Move it to 2 and you will see what happens 10ms later, and so on to 300ms (or whatever end point you choose for time range) - between 300ms and 600ms is best.

This shows how quickly (or not) your signal decays. If there are frequencies still ringing out at 600msec that aren't into the noise floor ~45dB, then you will definitely be still hearing these and the bass may be a bit muddy. You would like if the signal evenly decayed, fairly quickly into the noise.

Use an X=1, Y=100, Z=150 for general settings with the usual 45dB-105dB and 15Hz-200Hz axis...

Waterfall plots are very useful for frequencies below ~200Hz to analyze the modal response of your room.

brucek

**jvanhambelgium** · **07-31-08, 07:22 PM**

brucek, thanks for that usefull info.
I'll redo my measurement this weekend since I have my "measuring PC" back up-and-running...

I will try 2 main positions for the sub. The current one, and fully corner-loaded.(this means moving the sub against the wall for about 1.5meters)

I'll be posting some of them later...

Thanks so far!
Jan

Advertisement